Create or update query ruleset | Elasticsearch Guide

› › ›

Create or update query rulesetedit

This functionality is in technical preview and may be changed or removed in a future release. Elastic will apply best effort to fix any issues, but features in technical preview are not subject to the support SLA of official GA features.

Creates or updates a query ruleset.

Requestedit

PUT _query_rules/<ruleset_id>

Prerequisitesedit

Requires the manage_search_query_rules privilege.

(Required, object) Contains parameters for a query ruleset:

Request bodyedit

rules: (Required, array of objects) The specific rules included in this query ruleset.

Each rule must have the following information:

rule_id (Required, string) A unique identifier for this rule.
type (Required, string) The type of rule. At this time only pinned query rule types are allowed.
criteria (Required, array of objects) The criteria that must be met for the rule to be applied. If multiple criteria are specified for a rule, all criteria must be met for the rule to be applied.
actions (Required, object) The actions to take when the rule is matched. The format of this action depends on the rule type.

Criteria must have the following information:

type (Required, string) The type of criteria. The following criteria types are supported:
- exact Only exact matches meet the criteria defined by the rule. Applicable for string or numerical values.
- fuzzy Exact matches or matches within the allowed Levenshtein Edit Distance meet the criteria defined by the rule. Only applicable for string values.
- prefix Matches that start with this value meet the criteria defined by the rule. Only applicable for string values.
- suffix Matches that end with this value meet the criteria defined by the rule. Only applicable for string values.
- contains Matches that contain this value anywhere in the field meet the criteria defined by the rule. Only applicable for string values.
- lt Matches with a value less than this value meet the criteria defined by the rule. Only applicable for numerical values.
- lte Matches with a value less than or equal to this value meet the criteria defined by the rule. Only applicable for numerical values.
- gt Matches with a value greater than this value meet the criteria defined by the rule. Only applicable for numerical values.
- gte Matches with a value greater than or equal to this value meet the criteria defined by the rule. Only applicable for numerical values.
- always Matches all queries, regardless of input.
metadata (Optional, string) The metadata field to match against. Required for all criteria types except global.
values (Optional, array of strings) The values to match against the metadata field. Only one value must match for the criteria to be met. Required for all criteria types except global.

Actions depend on the rule type. For pinned rules, actions follow the format specified by the Pinned Query. The following actions are allowed:

ids (Optional, array of strings) The The unique document IDs of the documents to pin. Only one of ids or docs may be specified, and at least one must be specified.
docs (Optional, array of objects) The documents to pin. Only one of ids or docs may be specified, and at least one must be specified. You can specify the following attributes for each document:
- _index (Required, string) The index of the document to pin.
- _id (Required, string) The unique document ID.

Due to limitations within Pinned queries, you can only pin documents using ids or docs, but cannot use both in single rule. It is advised to use one or the other in query rulesets, to avoid errors. Additionally, pinned queries have a maximum limit of 100 pinned hits. If multiple matching rules pin more than 100 documents, only the first 100 documents are pinned in the order they are specified in the ruleset.

Examplesedit

The following example creates a new query ruleset called my-ruleset.

Two rules are associated with my-ruleset:

my-rule1 will pin documents with IDs id1 and id2 when user.query exactly matches marvel or dc and user.country exactly matches us.
my-rule2 will pin documents from different, specified indices with IDs id3 and id4 when the query_string fuzzily matches comic.

PUT _query_rules/my-ruleset
{
    "rules": [
        {
            "rule_id": "my-rule1",
            "type": "pinned",
            "criteria": [
                {
                    "type": "contains",
                    "metadata": "user.query",
                    "values": [ "marvel", "dc" ]
                },
                {
                    "type": "exact",
                    "metadata": "user.country",
                    "values": [ "us" ]
                }
            ],
            "actions": {
                "ids": [
                    "id1",
                    "id2"
                ]
            }
        },
        {
            "rule_id": "my-rule2",
            "type": "pinned",
            "criteria": [
                {
                    "type": "fuzzy",
                    "metadata": "query_string",
                    "values": [ "comic" ]
                }
            ],
            "actions": {
                "docs": [
                    {
                        "_index": "index1",
                        "_id": "id3"
                    },
                    {
                        "_index": "index2",
                        "_id": "id4"
                    }
                ]
            }
        }
    ]
}

« Query rules APIs Get query ruleset »